Building a Formative Evaluation Environment for Cooperative Learning Using Natural Language Processing and Learning Process Sensing

2023 Fiscal Year Final Research Report

• Project/Area Number: 19H01714
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 09070:Educational technology-related
• Research Institution: Shizuoka University
• Principal Investigator: Oshima Ritsuko 静岡大学, 情報学部, 教授 (70377729)
• Co-Investigator(Kenkyū-buntansha): 綱川 隆司 静岡大学, 情報学部, 講師 (30611214) ; 猿渡 俊介 大阪大学, 大学院情報科学研究科, 准教授 (50507811)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 協調学習 / 非言語データ / センシング技術 / 機械学習

Outline of Final Research Achievements

The purpose of this study was to develop an automatic grading system for learners' cooperative skills and to devise a method for analyzing group activities based on nonverbal behavior.
An automatic grading system, CRP, was built and equipped with functions for grading and teachers. To improve the accuracy of the machine learning algorithm for automatic grading of responses, we used BERT and BERTopic, and achieved a certain degree of grading accuracy. To analyze group activities based on nonverbal behaviors, we used SRP badges to collect and analyze physical movements during cooperative problem-solving tasks, and were able to obtain a general understanding of speaker detection, activity phase segmentation, and extraction of hot topics for discussion. We also evaluated the performance of the badges and algorithms, and found that the results were generally in line with our expectations. In addition, we constructed the SRP Web to visualize the collected data.

Free Research Field

学習科学

Academic Significance and Societal Importance of the Research Achievements

本研究の成果の意義は，協調学習支援研究において近年トピックとなっている社会共有的調整に着目し，学習者がその調整を行うために必要な協調スキルを，機械学習アルゴリズムを用いて自動評定する手法を検討した点にある．自動評定の実現は，教育現場において非常に有用性が高いだけでなく，今後多様な研究の評定手続きへの応用が見込まれ，社会科学研究への大きな貢献が期待できる．
また非言語行動と社会共有的調整の関係性は未だ解明されていない．学習者の協調スキルをセンサバッジで収集した非言語行動を元に解析し，言語データとの関係を明らかにすることで，協調学習における非言語行動の解明が進む．